• accelerometry
• AHA
• asymmetry index
• bimanual activity
• CP
• GMFCS.
• handedness
• HCP
• MACS
• upper limbs

This study aims to validate the triaxial accelerometer as a tool for measuring upper limb activity by exploring the relationship between the asymmetry index (i.e. the difference of the mean activity of each upper limb, obtained from the analysis of the accelerometer data), the Assisting Hand Assessment (AHA: a test of bimanual activity) score and the Manual Ability Classification System (MACS: a classification of the manual capability) level, in children with hemiplegic cerebral palsy (HCP). A systematic review of literature came first, exploring to what extent the amount of upper limbs activity of both healthy and children with disability can be reliably recorded and compared with each other, using a combination of multiple wearing inertial sensors (as accelerometers), in order to legitimize their use as feasible measurement in the experimental part of the study. The main conclusion of the review was that the use of inertial sensors worn simultaneously on both upper extremities is a valid tool for assessing upper motor activities and asymmetry, both in experimental and real-world conditions. For the clinical study, 14 children with HCP (9 male) and 8 control (6 males) were enrolled and assessed at mean age 9.74 years (standard deviation 3.22) at the Stella Maris Research Centre, Pisa - Italy and at the Queensland Cerebral Palsy and Rehabilitation Research Centre, Brisbane - Australia. Upper limb activity was assessed using two Actigraph triaxial accelerometers, worn on each wrist in two different situation: (i) while performing the Assisting Hand Assessment (AHA); and (ii) during a seven day record, performed with a diary completed for all daytime hours of weartime. Accelerometer data were collected at 80Hz and 1 second epochs, and summarized in Average Counts. The mean activity of the two upper limbs was analysed separately and compared with each other by non-parametric tests. Comparing the mean activity of the two hands of the same group, the dominant one showed higher values, in the both groups, but, as expected, it reached a statistically significant difference only in the HCP group (pTD= 0.75; pHCP< 0.0001). Accordingly with these results, the difference between the non dominant hands of the TD and HCP groups was statistically significant (p<0.0001), contrary to the comparison between the dominant hands (p=0.133). In addition, also the asymmetry index (AI) obtained for the HCP group was significantly higher (p<0.0001) than the one of the TD group. Moreover, comparing the MACS level and the AI during AHA, it was possible to distinguish 4 different clusters of upper limb motor ability (TD group, MACS I, II, and III groups). A further analysis was performed to investigate the relationship among data, using linear regression. The AI during AHA showed a high correlation with the AHA score (R2= .795; p<0.0001). The AI was calculated also considering the 7days data collection, for 16 (n=5 TD, n=11 HCP) participants of the whole sample, and it showed a high correlation with the AI during AHA (R2= .935; p<0.0000001). As a result, these data are very promising about the introduction of accelerometer approach as a reliable new objective and not invasive assessment tool for measuring the upper limb activity. It could become a new suitable common device to monitor, to evaluate and to tailor rehabilitation programs in children with HCP.